Reproduce “PersonaBOT: Bringing Customer Personas to Life with LLMs and RAG” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Muhammed Rizwan, Lars Carlsson, Mohammad Loni
Researchers from Jönköping University and Volvo Construction Equipment created PersonaBOT, a system leveraging Large Language Models and Retrieval-Augmented Generation to generate synthetic customer personas and make them interactively accessible. The system improved the chatbot's average accuracy rating for VCE stakeholders from 5.88 to 6.42 out of 10 after augmenting its knowledge base with LLM-generated personas.
Key insights
- Few-Shot prompting excelled in generating more complete customer personas, while Chain-of-Thought (CoT) prompting proved more efficient in terms of generation time and token usage.
- Systematically integrating LLM-generated synthetic personas into a RAG chatbot framework provides a scalable and interactive method for businesses to access customer insights.
- Iterative refinement of the RAG system, based on stakeholder feedback and augmented knowledge bases, is crucial for improving chatbot accuracy and aligning with business needs.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.